冷连轧过程控制系统及其先进控制算法的研究
|
摘要
现代化冷连轧生产的特点就是轧制速度的大幅度提升。在高速轧轧制的过程中，连轧机组的动态响应分析与控制相当重要，而动态控制的好坏很大程度上依赖于反映连轧动态过程的动态模型。本文结合思创电气工程有限公司“冷连轧机过程控制及其优化控制算法研究”科研项目，对冷连轧机动态模型及其控制算法进行了研究。
冷连轧机的自动控制是一种复杂的机-电-工艺的综合控制系统，变量较多，彼此相互影响、相互制约。而各机架参数又通过带钢，特别是张力相互联系起来。因此，一旦在某个机架上出现了调节量或干扰量的变化，不仅破坏了该机架的稳态，而且也会通过机架间的张力的变化传递给各个机架，从而使整个连轧机组的稳态遭到破坏，严重影响连轧机组的生产产量和生产质量。因此，基于改善连轧机组过程控制系统的复杂性，消除机架间的耦合作用，提高连轧机组轧制过程的稳定性，本文对相邻机架间的解耦控制进行了深入的研究，并针对冷连轧过程中的重复性，提出了一种迭代学习控制算法。
本文首先简述了带钢冷连轧计算机控制的发展现状，概述了迭代学习控制理论及其在国内外的应用现状。其次简述了冷连轧过程控制系统的组成，详细分析了ATC系统、AGC系统及AFC系统的基本结构和控制原理。然后给出了冷连轧轧制过程的基本参数和基本方程，在此基础上建立了冷连轧机基于稳态工作点的线性化动态模型，说明了相邻机架间各状态变量的耦合作用。以三连轧模型为控制对象，设计了最优解耦控制器，仿真结果从理论试验的角度验证了此方法的有效性。最后针对冷连轧过程的重复性、不确定性和高实时性，提出了一种迭代学习控制算法，并证明了此算法的收敛性，为迭代学习在轧钢领域的应用提供了一定的理论保证，同时也为迭代学习控制理论应用于实际工程进行了一次尝试。
High-speed rolling is one of the most important characteristics of modern cold tandem rolling mill.When the rolling mill runs fast, the dynamic response and the control of rolling mill are very important, and the result of dynamic control depends on the dynamic model which reflects the dynamic process of tandem rolling mill.Based on Strong Electrical Engineering Co. LTD. practical project, which is the "rolling process control system and optimization algorithm of cold tandem rolling mill control".This studay aims to investigate the dynamic model of rolling process and the control algorithm.
The autocontrol systems of tandem rolling mill are nonlinear, strong coupled complex systems with many variables.Every stand contact each other via a steel strip, especially with the tension.Hence, once the regulated variable or the disturbance of one stand varies, the stable rolling process of this stand will be destroyed, and the stable rolling process of other stands will also be destroyed via the variety of tension,so as to destroy stability of the whole rolling system.The yield and quality of rolling mills will be influenced seriously. In order to improve complexity of rolling process control system, to eliminate the interference between abutting stands, and also to improve the stability of the rolling process, this studay presents a optimal decoupling algorithm.Take the repetitive property of rolling process into account, an algorithm of ILC(Iterative Learning Control) of cold tandem rolling mill is proposed.
First, the computer control technique of cold tandem rolling mill is summarized, and the principle of ILC and the applied actuality of ILC domestic and oversea are introduced. Secondly, the components of cold tandem rolling process control system are introduced.This studay also analyzes the frame and the control theory of ATC(Automatic Tension Control) system, AGC(Automatic Gauge Control) system and AFC(Automatic Form Control) system particularly. Thirdly, the basic parameters and equations of rolling process are presented.A linearized dynamic model which illustrates the interference between abutting stands is presented. For instance the linearized dynamic models of cold tandem rolling mill with three stands, an optimal decoupling controller has been designed in this note, and the simulation result shows the effectiveness and correctness of the proposed approach.Considering the high real time property, repetitive property and uncertainty of cold tandem rolling process ,an algorithm of ILC is presented.This studay also proves the convergence of this algorithm.These provide theory basis for using ILC theory in cold tandem rolling process control. At the same time, this is a sample of using ILC theory in practical industrial systems.
冷连轧机的自动控制是一种复杂的机-电-工艺的综合控制系统，变量较多，彼此相互影响、相互制约。而各机架参数又通过带钢，特别是张力相互联系起来。因此，一旦在某个机架上出现了调节量或干扰量的变化，不仅破坏了该机架的稳态，而且也会通过机架间的张力的变化传递给各个机架，从而使整个连轧机组的稳态遭到破坏，严重影响连轧机组的生产产量和生产质量。因此，基于改善连轧机组过程控制系统的复杂性，消除机架间的耦合作用，提高连轧机组轧制过程的稳定性，本文对相邻机架间的解耦控制进行了深入的研究，并针对冷连轧过程中的重复性，提出了一种迭代学习控制算法。
本文首先简述了带钢冷连轧计算机控制的发展现状，概述了迭代学习控制理论及其在国内外的应用现状。其次简述了冷连轧过程控制系统的组成，详细分析了ATC系统、AGC系统及AFC系统的基本结构和控制原理。然后给出了冷连轧轧制过程的基本参数和基本方程，在此基础上建立了冷连轧机基于稳态工作点的线性化动态模型，说明了相邻机架间各状态变量的耦合作用。以三连轧模型为控制对象，设计了最优解耦控制器，仿真结果从理论试验的角度验证了此方法的有效性。最后针对冷连轧过程的重复性、不确定性和高实时性，提出了一种迭代学习控制算法，并证明了此算法的收敛性，为迭代学习在轧钢领域的应用提供了一定的理论保证，同时也为迭代学习控制理论应用于实际工程进行了一次尝试。
High-speed rolling is one of the most important characteristics of modern cold tandem rolling mill.When the rolling mill runs fast, the dynamic response and the control of rolling mill are very important, and the result of dynamic control depends on the dynamic model which reflects the dynamic process of tandem rolling mill.Based on Strong Electrical Engineering Co. LTD. practical project, which is the "rolling process control system and optimization algorithm of cold tandem rolling mill control".This studay aims to investigate the dynamic model of rolling process and the control algorithm.
The autocontrol systems of tandem rolling mill are nonlinear, strong coupled complex systems with many variables.Every stand contact each other via a steel strip, especially with the tension.Hence, once the regulated variable or the disturbance of one stand varies, the stable rolling process of this stand will be destroyed, and the stable rolling process of other stands will also be destroyed via the variety of tension,so as to destroy stability of the whole rolling system.The yield and quality of rolling mills will be influenced seriously. In order to improve complexity of rolling process control system, to eliminate the interference between abutting stands, and also to improve the stability of the rolling process, this studay presents a optimal decoupling algorithm.Take the repetitive property of rolling process into account, an algorithm of ILC(Iterative Learning Control) of cold tandem rolling mill is proposed.
First, the computer control technique of cold tandem rolling mill is summarized, and the principle of ILC and the applied actuality of ILC domestic and oversea are introduced. Secondly, the components of cold tandem rolling process control system are introduced.This studay also analyzes the frame and the control theory of ATC(Automatic Tension Control) system, AGC(Automatic Gauge Control) system and AFC(Automatic Form Control) system particularly. Thirdly, the basic parameters and equations of rolling process are presented.A linearized dynamic model which illustrates the interference between abutting stands is presented. For instance the linearized dynamic models of cold tandem rolling mill with three stands, an optimal decoupling controller has been designed in this note, and the simulation result shows the effectiveness and correctness of the proposed approach.Considering the high real time property, repetitive property and uncertainty of cold tandem rolling process ,an algorithm of ILC is presented.This studay also proves the convergence of this algorithm.These provide theory basis for using ILC theory in cold tandem rolling process control. At the same time, this is a sample of using ILC theory in practical industrial systems.
引文
1 潘学军，张承进，柴天佑．冷连轧过程线性化动态模型的研究[J]．系统仿真学报 1997，10(6)：35．
2 Foster M. A., Marshell S..A. Modelling of a tandem cold strip rolling mill[J].Mathematicai process models in iron and steel making , Amsteerdam. Feburary, 1973: 19-21.
3 杨杰．轧制过程数学模型[M]．北京：冶金工业出版社，1993．
4 薛兴昌．轧钢自动化系统的功能评述和技术发展[J]．钢铁．1997，32(增刊)：880-884．
5 孙明轩，黄宝健．迭代学习控制[M]．北京：国防工业出版社，1999．
6 谢胜利，田森平，谢振东．迭代学习控制的理论与应用[M]．北京：科学出版社，2005．
7 华建新，王贞祥．全连续式冷连轧机过程控制[M]．北京：冶金工业出版，2000．
8 李庆尧，杨弘鸣．带钢冷连轧机过程控制计算机及应用软件设计[M]．北京：冶金工业出版社，1995．
9 王国栋，刘相华，王军生．冷连轧生产工艺的进展[J]．轧钢．2003，20(1)：37-41．
10 张进之，石勇．冷连轧机控制系统及智能控制系统的开发[J]．重型机械．2001，14-18．
11 孙一康．带钢冷连轧计算机控制[M]．北京：冶金工业出版社，2002．
12 王国栋，刘相华，王军生．冷连轧计算机过程控制系统[J]．轧钢．2003，20(2)：41-44．
13 Uchiyama M. Formation of high-speed motion pattern of a mechnical arm by trial[J]. ransaction of the Society of Instrumentration and Control Engineers. 1978, 14(6): 706-712.
14 Arimoto S, Kawamura S, Miyazaki F. Bettering operation of robots by learning[J]. Journal or Robotic Systems. 1984, 1(2): 123-140.
15 Arimoto S, Kawamura S, Miyazaki F. Bettering operation of dynamic systems by learning: A new control theory for servomechanism or mechatronics systems[J]. In: Proceedings of the 23rd IEEE Conference on Decision and Control, LasVesgas, Nevada. 1984, 1064-1069.
16 Arimoto S. Learning control theory for robot motion [J]. Int.J.Adapt. Contr.Signal Processing. 1990(4): 543-564.
17 Guoli, HaoLi, Jian Xin Xu. The study and design of the longitudinal Icarning control of the automobile[J]. In Proceedings of the 2-end Asian Control Conference, Seoul, Korea, July, 1997.
18 Chien ChongChen,Chyi Hwang, Ray Y. K. Yang. Optimal periodirc forcing of nonlinear chemical processed for performance improvements[J]. The Canadian Journal of ChemicalEngineering. August, 1994, 72: 672-682.
19 张乐．提高CNC活塞加工精度的研究[D]．西安：西安交通大学，1994．
20 Geng Z, Carroll R, Janmshidi M.An. Adaptive Learning Control Approach with Application to Water Tank Level Control[J]. Control Theory Advanced Tech. 1992, 8(3): 577-592.
21 Pandit M, Buchheit K. H. Optimizing Iterative Learning Control of Cyclic Production Process with application to extruders[J]. IEEE Trans Contr Tech. 1999, 7(3): 382-390.
22 Srinivas S.Garimella, Krishnaswamy (Cheena) Srinivasa. Application of Iterative Learning Control to Coil-to-Coil in Rolling[J]. IEEE Transaction On Control Systems Technology. March, 1998, 6(2): 281-293.
23 陈哲名，潘再平．基于迭代学习的开关磁阻电动机最优控制研究[J]．浙江大学学报．2006，40(1)：17-22．
24 王东兴，侯明，胡平平．变频调速系统离散化迭代学习控制法[J]．电工技术学报．2005，20(5)：67-71．
25 曹阳，沈毅力，李天石．加速闭环P型迭代学习在电液伺服协调加载系统中的应用[J]．机床与液压．2000，(5)：38-39．
26 Cheng Shao, Rong Gao Fu, Yang Yi. Robust stability of Optimal Iterative Learning Control and application to injection molding machine[J]. ActaAutomatic. 2003, 29(1): 73-79.
27 骆传刚，赵弘等．迭代学习法在机械压力机滑块位置控制中的应用[J]．机床与液压．2003(5)：31-33．
28 赵升吨，宋涛等．迭代学习法在开关式液压阀位置控制中的应用[J]．西安交通大学学报．2003，37(5)：36-41．
29 刘伟，秦伟，王蜀霞．液压同步运动系统自适应学习控制策略和方法[J]．中国机械工程．1996，7(1)：131-136．
30 唐谋凤．现代带钢冷连轧的自动化[M]．北京：冶金土业出版，1995．
31 汪祥能．现代带钢连轧机控制[M]．沈阳：东北大学出版社，1996．
32 Dendle D. W. A. Review of Automatic Control Systems for Cold Tandem Mills[J]. Steel Time. 1979, 207(3): 78-85.
33 贺敏辛．冷轧板带生产[M]．冶金工业出版社，北京：1990．
34 徐光，陈和铁，杨节．冷连轧机控制系统[J]．冶金自动化．1999，(2)：38-40．
35 Bryant. Automation of tandem mill[J]. The metals society publication, 1973.
36 Watanaba H, Takahaski N. Control Equations for Dynamic Characteristics of Cold Rolling Tandem Mills[J] . Iron and Steel Engineer Year Book. 1974: 91-96.
37 刘长玉，魏丽．带钢冷连轧机架间张力调节器的合理结构[J]．冶金自动化．1992，16(2)：19-22．
38 刘建昌，钱晓龙，陈宏志．冷连轧机张力自动控制系统[J]．钢铁．2002，37(12)：36-39．
39 连家创，刘宏民．板厚板形控制[M]．北京：兵器业出版社，1996．
40 Kazuma Gumi. New gage control system for tandem cold mill Iron and Steel Engineer. 1994, 12.
41 王国栋．板形控制和板形理论[M]．北京：冶金工业出版社，1986．
42 赵志业．金属塑性变形与轧制理论[M]．北京：冶金工业出版社，1994．
43 郭志强．整体机架刚度特性的有限元法研究[J]．包头钢铁学院学报．1994，13(1)：49—57．
44 张树堂，刘玉荣．变断面张力微分方程与冷连轧动态数学模拟数学模型[J]．金属学报．1981，17(4)：206-212．
45 张进之．连轧张力公式[J]．金属学报．1978，14(6)：127-138．
46 马文博，徐光，杨永立．冷连轧张力公式推导[J]．钢铁研究．2003，3(132)：28-31．
47 Arimura T. An Analysis into the Dynamic-Behaviour of Tandem Cold Mills[J]. Proceedings ICSTIS, Suppl, Trans, ISIJ. 1971,, (11): 777-781.
48 吴国发，徐哲，卓兵．冷连轧系统跟踪问题的最优控制[J]．冶金自动化．1986，(6)：36-41．
49 王修贤，乔辉，李军忠．基于线性化模型的冷连轧机PI解耦控制系统[J]．工业仪表与自动化装置，2004(4)：3-5．
50 Yasunori Katayama, Mito, Yasuo Morooka, Taiko Takano. Non-interference control method and device[J]. United States Patent Application Publication, US5375448．Dec．27，1994．
51 张兴国，林辉．迭代学习控制理论进展与展望[J]．测控技术．2006，25(11)：1-5．
52 D. H. Owens, J.atonen. Iterative learning control-An optimization paradigm[J]. Annulal Reviews in control. 2005(29): 57-70.
53 许顺孝，杨富文．不确定线性系统迭代学习控制器的设计[J]．控制理论与应用．2002，19(4)：650-652．
54 吴杰，许忠仁，詹习生，李铁春．不确定线性鲁棒迭代学习控制其设计及应用[J]．自动化与仪表．2005(5)：(1-4)．
55 孙明轩．迭代学习控制系统收敛性与鲁棒性[J]．西北工业学院学报．1996，1(16)：1-7．
56 孙明轩．迭代学习控制系统的鲁棒性分析[J]．科技通报．1996，12(4)：198-203．
57 王建国，阮小娥．工业过程迭代学习算法的鲁棒性[J]．工程数学学报．2004，21(1)：127-130．
58 刘山，吴铁军．基于H∞方法的鲁棒迭代学习控制设计[J]．浙江大学学报．2004，38(3)：257-261．
59 Roel Merry, Renevande Molengraft, Maarten Steinbuch. Disturbances and model uncertainties in iterative learning control[J]. The Fourth International Workshop On Multidimensional System, 2005.
60 TAYEBI A, ZAREMBA M. B. Robust iterative learning control design is straightforward for uncertain LTI systems satisfying the robust performance condition[J]. IEEE Trans Automatic Control. 2003, 48(1): 101-106.
2 Foster M. A., Marshell S..A. Modelling of a tandem cold strip rolling mill[J].Mathematicai process models in iron and steel making , Amsteerdam. Feburary, 1973: 19-21.
3 杨杰．轧制过程数学模型[M]．北京：冶金工业出版社，1993．
4 薛兴昌．轧钢自动化系统的功能评述和技术发展[J]．钢铁．1997，32(增刊)：880-884．
5 孙明轩，黄宝健．迭代学习控制[M]．北京：国防工业出版社，1999．
6 谢胜利，田森平，谢振东．迭代学习控制的理论与应用[M]．北京：科学出版社，2005．
7 华建新，王贞祥．全连续式冷连轧机过程控制[M]．北京：冶金工业出版，2000．
8 李庆尧，杨弘鸣．带钢冷连轧机过程控制计算机及应用软件设计[M]．北京：冶金工业出版社，1995．
9 王国栋，刘相华，王军生．冷连轧生产工艺的进展[J]．轧钢．2003，20(1)：37-41．
10 张进之，石勇．冷连轧机控制系统及智能控制系统的开发[J]．重型机械．2001，14-18．
11 孙一康．带钢冷连轧计算机控制[M]．北京：冶金工业出版社，2002．
12 王国栋，刘相华，王军生．冷连轧计算机过程控制系统[J]．轧钢．2003，20(2)：41-44．
13 Uchiyama M. Formation of high-speed motion pattern of a mechnical arm by trial[J]. ransaction of the Society of Instrumentration and Control Engineers. 1978, 14(6): 706-712.
14 Arimoto S, Kawamura S, Miyazaki F. Bettering operation of robots by learning[J]. Journal or Robotic Systems. 1984, 1(2): 123-140.
15 Arimoto S, Kawamura S, Miyazaki F. Bettering operation of dynamic systems by learning: A new control theory for servomechanism or mechatronics systems[J]. In: Proceedings of the 23rd IEEE Conference on Decision and Control, LasVesgas, Nevada. 1984, 1064-1069.
16 Arimoto S. Learning control theory for robot motion [J]. Int.J.Adapt. Contr.Signal Processing. 1990(4): 543-564.
17 Guoli, HaoLi, Jian Xin Xu. The study and design of the longitudinal Icarning control of the automobile[J]. In Proceedings of the 2-end Asian Control Conference, Seoul, Korea, July, 1997.
18 Chien ChongChen,Chyi Hwang, Ray Y. K. Yang. Optimal periodirc forcing of nonlinear chemical processed for performance improvements[J]. The Canadian Journal of ChemicalEngineering. August, 1994, 72: 672-682.
19 张乐．提高CNC活塞加工精度的研究[D]．西安：西安交通大学，1994．
20 Geng Z, Carroll R, Janmshidi M.An. Adaptive Learning Control Approach with Application to Water Tank Level Control[J]. Control Theory Advanced Tech. 1992, 8(3): 577-592.
21 Pandit M, Buchheit K. H. Optimizing Iterative Learning Control of Cyclic Production Process with application to extruders[J]. IEEE Trans Contr Tech. 1999, 7(3): 382-390.
22 Srinivas S.Garimella, Krishnaswamy (Cheena) Srinivasa. Application of Iterative Learning Control to Coil-to-Coil in Rolling[J]. IEEE Transaction On Control Systems Technology. March, 1998, 6(2): 281-293.
23 陈哲名，潘再平．基于迭代学习的开关磁阻电动机最优控制研究[J]．浙江大学学报．2006，40(1)：17-22．
24 王东兴，侯明，胡平平．变频调速系统离散化迭代学习控制法[J]．电工技术学报．2005，20(5)：67-71．
25 曹阳，沈毅力，李天石．加速闭环P型迭代学习在电液伺服协调加载系统中的应用[J]．机床与液压．2000，(5)：38-39．
26 Cheng Shao, Rong Gao Fu, Yang Yi. Robust stability of Optimal Iterative Learning Control and application to injection molding machine[J]. ActaAutomatic. 2003, 29(1): 73-79.
27 骆传刚，赵弘等．迭代学习法在机械压力机滑块位置控制中的应用[J]．机床与液压．2003(5)：31-33．
28 赵升吨，宋涛等．迭代学习法在开关式液压阀位置控制中的应用[J]．西安交通大学学报．2003，37(5)：36-41．
29 刘伟，秦伟，王蜀霞．液压同步运动系统自适应学习控制策略和方法[J]．中国机械工程．1996，7(1)：131-136．
30 唐谋凤．现代带钢冷连轧的自动化[M]．北京：冶金土业出版，1995．
31 汪祥能．现代带钢连轧机控制[M]．沈阳：东北大学出版社，1996．
32 Dendle D. W. A. Review of Automatic Control Systems for Cold Tandem Mills[J]. Steel Time. 1979, 207(3): 78-85.
33 贺敏辛．冷轧板带生产[M]．冶金工业出版社，北京：1990．
34 徐光，陈和铁，杨节．冷连轧机控制系统[J]．冶金自动化．1999，(2)：38-40．
35 Bryant. Automation of tandem mill[J]. The metals society publication, 1973.
36 Watanaba H, Takahaski N. Control Equations for Dynamic Characteristics of Cold Rolling Tandem Mills[J] . Iron and Steel Engineer Year Book. 1974: 91-96.
37 刘长玉，魏丽．带钢冷连轧机架间张力调节器的合理结构[J]．冶金自动化．1992，16(2)：19-22．
38 刘建昌，钱晓龙，陈宏志．冷连轧机张力自动控制系统[J]．钢铁．2002，37(12)：36-39．
39 连家创，刘宏民．板厚板形控制[M]．北京：兵器业出版社，1996．
40 Kazuma Gumi. New gage control system for tandem cold mill Iron and Steel Engineer. 1994, 12.
41 王国栋．板形控制和板形理论[M]．北京：冶金工业出版社，1986．
42 赵志业．金属塑性变形与轧制理论[M]．北京：冶金工业出版社，1994．
43 郭志强．整体机架刚度特性的有限元法研究[J]．包头钢铁学院学报．1994，13(1)：49—57．
44 张树堂，刘玉荣．变断面张力微分方程与冷连轧动态数学模拟数学模型[J]．金属学报．1981，17(4)：206-212．
45 张进之．连轧张力公式[J]．金属学报．1978，14(6)：127-138．
46 马文博，徐光，杨永立．冷连轧张力公式推导[J]．钢铁研究．2003，3(132)：28-31．
47 Arimura T. An Analysis into the Dynamic-Behaviour of Tandem Cold Mills[J]. Proceedings ICSTIS, Suppl, Trans, ISIJ. 1971,, (11): 777-781.
48 吴国发，徐哲，卓兵．冷连轧系统跟踪问题的最优控制[J]．冶金自动化．1986，(6)：36-41．
49 王修贤，乔辉，李军忠．基于线性化模型的冷连轧机PI解耦控制系统[J]．工业仪表与自动化装置，2004(4)：3-5．
50 Yasunori Katayama, Mito, Yasuo Morooka, Taiko Takano. Non-interference control method and device[J]. United States Patent Application Publication, US5375448．Dec．27，1994．
51 张兴国，林辉．迭代学习控制理论进展与展望[J]．测控技术．2006，25(11)：1-5．
52 D. H. Owens, J.atonen. Iterative learning control-An optimization paradigm[J]. Annulal Reviews in control. 2005(29): 57-70.
53 许顺孝，杨富文．不确定线性系统迭代学习控制器的设计[J]．控制理论与应用．2002，19(4)：650-652．
54 吴杰，许忠仁，詹习生，李铁春．不确定线性鲁棒迭代学习控制其设计及应用[J]．自动化与仪表．2005(5)：(1-4)．
55 孙明轩．迭代学习控制系统收敛性与鲁棒性[J]．西北工业学院学报．1996，1(16)：1-7．
56 孙明轩．迭代学习控制系统的鲁棒性分析[J]．科技通报．1996，12(4)：198-203．
57 王建国，阮小娥．工业过程迭代学习算法的鲁棒性[J]．工程数学学报．2004，21(1)：127-130．
58 刘山，吴铁军．基于H∞方法的鲁棒迭代学习控制设计[J]．浙江大学学报．2004，38(3)：257-261．
59 Roel Merry, Renevande Molengraft, Maarten Steinbuch. Disturbances and model uncertainties in iterative learning control[J]. The Fourth International Workshop On Multidimensional System, 2005.
60 TAYEBI A, ZAREMBA M. B. Robust iterative learning control design is straightforward for uncertain LTI systems satisfying the robust performance condition[J]. IEEE Trans Automatic Control. 2003, 48(1): 101-106.